July 1, 2020, 16:42 (GMT) |
Change boolean blenlib interface to be purely C++. |
July 1, 2020, 13:15 (GMT) |
Change bmesh_boolean.c into a C++ file and start using PolyMesh interface. Also, use BLI_math_mpq.hh everywhere instead of gmpxx.h. |
June 30, 2020, 19:31 (GMT) |
Fix problem after master merge with Optional -> optional migration. On Mac, there's an error re using the value() member of std::optional. Working around it for now. |
June 30, 2020, 18:54 (GMT) |
Merge branch 'master' into newboolean |
June 30, 2020, 18:46 (GMT) |
Dissolving triangulation edges and verts passes first test. |
June 28, 2020, 22:47 (GMT) |
First test with dissolve of triangulation edges is passing. Still lots to do. And this includes a horribly inefficient way of finding which edges are dissolvable -- to be fixed later. |
Revision ce37092 by Howard Trickey (newboolean) June 26, 2020, 12:36 (GMT) |
Merge branch 'master' into newboolean |
Revision f80378c by Howard Trickey (newboolean) June 19, 2020, 19:53 (GMT) |
Continued work on getting triangulation edges removed from output. |
Revision 59805fa by Howard Trickey (newboolean) June 17, 2020, 12:58 (GMT) |
Merge branch 'master' into newboolean |
Revision d408199 by Howard Trickey (newboolean) June 17, 2020, 12:55 (GMT) |
Start of PolyMesh interface for boolean. Also refactored delaunay to put the 2d mesh representation into an externally visible CDTArrangement class. Not sure yet whether I will use this; if I end up not using it, will move that part of the header back into the implementation file. |
June 14, 2020, 11:58 (GMT) |
Fixed bug for some binary boolean ops. Fixed a memory leak. Now the basic binary booleans work from the tool (not the modifier). Still to do: (1) coplanar and non-connected-component cases; (2) dissolve triangulation lines and new verts due only to them; (3) properly transfer mesh attributes from old to new; (4) hook up to modifier; (5) attention to and tuning of performance. |
June 13, 2020, 23:54 (GMT) |
Made binary version of boolean and hooked up to Boolean tool. Union seems to work. Other ops are flaky because I haven't quite got the winding number propagation right yet. |
June 13, 2020, 15:11 (GMT) |
More boolean tests pass. Fixed bug re confusion of how to sort triangles around an edge. Fixed bug in first tettet test (inconsistent normals in input). |
June 12, 2020, 14:50 (GMT) |
Proper implementation of finding ambient cell. Also fixed bug in boolean test where passed wrong output to the obj writer. |
June 12, 2020, 00:37 (GMT) |
First real boolean union output test works. |
June 11, 2020, 22:46 (GMT) |
Quient some warnings on Linux. |
June 11, 2020, 17:47 (GMT) |
After merge from master, updated for changes to BLI. Mainly namespace changes, BLI -> blender, so I changed my own namespaces to follow that pattern. ArrayRef -> Span. Now can have hash() as member of data structures rather that having to make a DefaultHash for them. Fixed a bad assumption about Arrays of bool being initialized to false. |
June 11, 2020, 09:44 (GMT) |
Merge branch 'master' into newboolean |
June 11, 2020, 09:41 (GMT) |
Format previous commit. |
June 11, 2020, 02:00 (GMT) |
A lot of progress towards working boolean library function. The code to partition space into cells is mostly done. The code to propagate winding numbers and flag cells according to the boolean function is mostly done. Only final extraction code is left, and a few other things. |
|